A Systematic Review on the Application of AI in the Enhancement of Control and Optimisation of Grid-Tied Microgrids

Abstract

Grid-tied Microgrids (MGs) are gaining popularity as a solution to the increasing demand for reliable and affordable energy. Integration of these MGs into the grid is challenging due to the fluctuating nature of Renewable Energy Sources (RESs), which is coupled with grid instability, faults and changing loads. Traditional control and optimisation methods cannot fully address the dynamic requirements of controlling these grid-tied MGs. In recent years, Artificial Intelligence (AI) techniques have emerged as a viable approach to improve the control and optimisation of MGs. This paper reviews grid-tied MGs, focusing on renewable energy integration, control and optimisation strategies. The methodology for the review process was preceded by a structured identification of a sample size of 312 relevant studies from major scientific databases, screening based on predefined inclusion and exclusion criteria, eligibility assessment, and final inclusion of high-quality peer-reviewed articles that were published between 2000 and 2025. The objective of this study was to analyse the evolution and performance of various types of microgrid control strategies and optimisation methods, with emphasis on AI techniques like Machine Learning and Reinforcement Learning and specifically on hybrid control architectures which combine model-based approaches with data-driven intelligence to enhance system stability and adaptability. Reviewed sources indicate that AI-controlled microgrid systems improve the performance of traditional control systems when confronted with non-linear dynamics and uncertainties in renewable energy resources, as well as load fluctuations. The results of the analysis also indicate that AI-based control systems contribute to improved power quality, regulation and efficiency through analysing total harmonic distortion and voltage/frequency regulation. Additionally, the emerging trends in predictive maintenance and fault detection are identified as key contributors to the reliability and resilience of these networks. AI control and optimisation methodologies create a comprehensive framework for the control and optimisation of the next-generation grid-tied microgrid, thereby supporting the transition to intelligent, sustainable, and decentralised energy systems.